A gas turbine surge margin early warning method and system

Abstract

The application discloses a gas turbine compressor surge boundary early warning method and system, and belongs to the technical field of gas turbine safety control. The method is based on the law of conservation of mass and momentum, establishes a nonlinear dynamics model of the compressor, the cavity and the throttle valve system, and constructs a cubic nonlinear stiffness oscillation system with dynamically updated natural frequency. The key parameters of the model are identified online, the inherent high-frequency disturbance of the system is extracted as auxiliary excitation, and the response amplitude R of the characteristic low-frequency surge precursor mode is calculated. By establishing an R value baseline model, the abnormal rising trend thereof is monitored to trigger an ultra-early warning. The system comprises data acquisition, model identification, response monitoring and warning output modules, can be integrated in a controller or a health management unit and can drive an active control system. The application does not require additional sensors, can realize high-sensitivity and high-robustness warning in a wide operating condition range, and provides a decision basis for active stability control of an aero-engine.

Description

Technical Field

[0001] This invention relates to the field of gas turbine safety control technology, specifically to a gas turbine surge boundary early warning method and system. Background Technology

[0002] Compressor surge is the most dangerous aerodynamic and thermodynamic instability phenomenon in gas turbines, characterized by low-frequency, large-amplitude oscillations of airflow along the compressor axis. It can lead to a sudden drop in thrust, component damage, and even scrapping. Traditional surge warning methods mainly rely on monitoring the root mean square value of the compressor outlet total pressure or identifying precursory features such as rotating stall. These are outcome-based monitoring methods, which suffer from problems such as short warning times and insensitivity to early surge precursor modes, making it difficult to provide sufficient intervention time for the control system.

[0003] Existing stability theories are mostly based on linear or weakly nonlinear assumptions, treating surge boundaries as static lines, and thus failing to reflect the dynamic stability changes of the system under real-time disturbances. In recent years, nonlinear dynamics theory has provided a new approach for detecting the early response of systems in complex environments, namely, utilizing the interaction between high-frequency auxiliary excitation and the nonlinear system to amplify the response to low-frequency information. Research shows that the instability process of a compressor system can be abstracted as a nonlinear oscillator, and its dynamic behavior has a profound intrinsic connection with the equations of a nonlinear oscillating system. Summary of the Invention

[0004] The purpose of this invention is to provide a gas turbine surge boundary early warning method and system to solve the problems of short window period and inability to detect changes in system dynamic stability in existing surge early warning methods.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: A gas turbine surge boundary early warning method includes the following steps: S1: Based on the laws of conservation of mass and momentum, establish the dimensionless control equations of the compressor-cavity-throttle valve system and construct the nonlinear dynamic model of the compressor system; S2: Fit the compressor characteristic curve, introduce disturbance variables to construct a nonlinear oscillating system model with cubic nonlinear stiffness. The natural frequency of the nonlinear oscillating system model can be dynamically updated. The equations of the nonlinear oscillating system model are: In the formula ζ represents the dimensionless flow disturbance or pressure pulsation; ζ represents the system's equivalent damping ratio; and η represents the nonlinear stiffness coefficient. This is the system's inherent frequency; For system incentives; The system's natural frequency The dynamic update rule expression is: The expression for the system excitation F(t) is: In the formula, The learning rate; The characteristic low-frequency frequency of the surge precursor mode; To characterize the characteristic low-frequency information of surge precursor modes; This is the system's inherent high-frequency frequency; The high-frequency information is inherent to the system and serves as an auxiliary stimulus; For system uncertainty disturbances; S3: Based on compressor characteristic data and operating data, identify the nonlinear stiffness coefficient η, damping ratio ζ, and natural frequency of the nonlinear oscillation system model online. ; S4: Collect compressor operating parameters and extract high-frequency disturbance information from the collected parameters. Calculate the nonlinear oscillation system at the characteristic low frequency. The response amplitude R at the point is used to establish a baseline model of how the R value changes with time or operating conditions. When the R value exceeds the baseline threshold and continues to rise, it is determined that the system is approaching the surge boundary and an early warning is issued. A complete technical chain of "modeling-identification-monitoring-early warning" is constructed. Based on nonlinear dynamics theory, the compressor system is abstracted into an oscillating system. The inherent high-frequency disturbance is used to amplify the low-frequency surge precursor signal, and the early warning is achieved through quantitative monitoring of the response amplitude. The core technical solutions are covered to the greatest extent, realizing ultra-early surge early warning under a wide range of operating conditions, and providing sufficient decision time for active stability control.

[0006] In a further embodiment, in step S2, the key parameters of the nonlinear oscillating system model with cubic nonlinear stiffness are related to the compressor's physical characteristic parameters through the following formula: In the formula, For stability parameters, It is related to rotational speed, cavity volume, and flow channel size; β is the linearization coefficient of the throttle valve; This is the first derivative of the compressor characteristic curve at the steady operating point; This is the second derivative of the compressor characteristic curve at the steady operating point; This is the third derivative of the compressor characteristic curve at the steady operating point; A function characterizing the relationship between natural frequency and compressor physical characteristic parameters; It is a function that characterizes the relationship between the damping ratio and the physical characteristic parameters of the compressor; It is a function that characterizes the relationship between the nonlinear stiffness coefficient and the physical characteristic parameters of the compressor; Establish a precise mapping relationship between mathematical model parameters and the actual physical characteristics of the compressor, and open up the connection between theoretical models and engineering practice; ensure that model parameters have clear physical meaning, improve the model's ability to represent the real dynamic characteristics of the compressor, and enhance the reliability of early warning methods.

[0007] In a further embodiment, the specific process of constructing a nonlinear oscillating system model with cubic nonlinear stiffness in step S2 is as follows: at the stable operating point of the compressor, a third-order Taylor expansion of the compressor characteristic curve is performed, and the second-order nonlinear term is eliminated by coordinate transformation to obtain a nonlinear oscillating system model describing the axial reciprocating oscillation characteristics of the compressor airflow. The analysis process of complex aerodynamic systems is simplified by mathematical transformations, while retaining the third-order nonlinear terms that play a key role in system instability. The nonlinear instability behavior of the compressor is accurately captured, providing a concise and effective model basis for subsequent response amplitude calculations.

[0008] In a further embodiment, the parameter identification process in step S3 includes the following stages: First stage: Based on the steady-state characteristic curve data of the target compressor, and according to the real-time gas turbine speed and throttle valve opening, determine the current stable operating point. ,in To stabilize the dimensionless flow coefficient at the operating point, The dimensionless pressure coefficient at the steady-state operating point; Second stage: Calculate the compressor characteristic curve at the steady-state operating point. The first derivative at Second derivative Third derivative The first stage involves calculating the linearization coefficient β of the throttle valve under the current operating conditions and the stability parameter B. The second stage, based on the calculated compressor physical characteristic parameters, involves obtaining the nonlinear stiffness coefficient η, damping ratio ζ, and natural frequency. ; A phased, process-oriented parameter identification strategy is adopted, which gradually transforms operating data into physical parameters and then into model parameters; this improves the efficiency and accuracy of parameter identification and ensures that the model can adapt to changes in the operating conditions of the gas turbine in real time.

[0009] In a further embodiment, in step S4, the nonlinear oscillation system processes characteristic low-frequency information. The response amplitude R is calculated using the following formula: In the formula, Characteristic low frequency The sine Fourier coefficient at the location; Characteristic low frequency The cosine Fourier coefficient at the point; and By analyzing the steady-state response time series of the system model Obtained by integration or discrete summation; Fourier analysis is used to quantitatively extract the precursor response signal of low-frequency surge, establish an intuitive indicator for determining stability margin, provide accurate quantitative early warning basis, avoid subjective judgment errors, and improve the consistency and repeatability of early warning methods.

[0010] In a further embodiment, in step S4, the compressor operating parameters include compressor outlet pressure, outlet flow rate, rotor speed, and inlet condition parameters. This clarifies the data source range required by the method and limits the specific type of parameter acquisition. No additional dedicated sensors are required; existing gas turbine monitoring parameters can be directly utilized, reducing engineering implementation costs and integration difficulty.

[0011] In a further embodiment, in step S4, the high-frequency disturbance information The source of the noise is at least one of the combustion noise inside the gas turbine system and the blade passing frequency. The system's own inherent high-frequency signal is used as an auxiliary excitation, without the need for additional external excitation devices. This simplifies the implementation conditions of the method, avoids interference from external excitation on the normal operation of the system, and improves the practicality of the engineering.

[0012] In a further embodiment, in step S4, the process of establishing the baseline model is as follows: during the healthy and stable operation phase of the gas turbine, the R value is collected under different speeds and different throttle valve openings, a correlation model between the R value and the operating parameters is established, and a reference benchmark for early warning judgment is constructed based on the full operating condition health data to achieve adaptive threshold setting for operating conditions; this enables the early warning judgment to have an objective and unified standard, avoids the limitations of a single threshold under a wide range of operating conditions, and improves the accuracy and robustness of the early warning.

[0013] A gas turbine surge boundary early warning system for implementing the above method includes: a data acquisition module for online acquisition of compressor operating parameters and extraction of high-frequency disturbance information from the acquired parameters. Model identification module: used to derive the nonlinear dynamic model of the compressor system and identify and update the parameters of the nonlinear oscillating system with cubic nonlinear stiffness; Response monitoring module: used to calculate the system response amplitude R online and perform trend analysis and baseline comparison on the R value; Early warning output module: used to generate graded early warning information when the R value exceeds the baseline threshold and continues to rise, and output the early warning information to the gas turbine control system. The modular system architecture is designed according to the technical steps of the method to realize the engineering implementation of the method; The modular structure facilitates the maintenance, upgrading and functional expansion of the system, forming a complete technical solution with the early warning method.

[0014] In a further embodiment, the system is integrated into the gas turbine controller or a separate health management unit. The system supports linkage with the gas turbine active stability control system. The sampling frequency of the data acquisition module is matched with the inherent high frequency of the compressor. The system's deployment method, linkage capability, and key technical details are clearly defined. This improves the compatibility between the system and the gas turbine, ensures the accuracy of data acquisition, supports the rapid conversion of early warning information into control commands, and realizes closed-loop management of "early warning-control".

[0015] The present invention has the following beneficial effects: This invention combines nonlinear dynamics theory with compressor aerodynamic stability monitoring, and uses the system's inherent high-frequency disturbance to amplify the response signal of the low-frequency surge precursor mode. This allows the system to capture the trend of decreasing stability margin before surge actually occurs, thereby extending the early warning time window and providing sufficient intervention time for active stability control. The dynamic learning mechanism of the inherent frequency enables the method to adapt to a wide range of speeds and operating conditions, breaking through the limitation of traditional methods that are only applicable to specific operating conditions. The adoption of simple and effective learning rules avoids the system instability caused by complex algorithms, ensuring the robustness and reliability of the early warning process. At the same time, the method directly uses data collected by conventional sensors at the compressor outlet, without the need for additional special equipment, reducing engineering integration costs and possessing high practical value. Attached Figure Description

[0016] Figure 1 : Overall flowchart of the method of the present invention.

[0017] Figure 2 : Comparison of the early warning effects of the method of this invention and the traditional method in a simulated surge event.

[0018] Figure 3 The diagram illustrates the specific steps of the method of the present invention. Detailed Implementation

[0019] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0020] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. These embodiments strictly follow the technical process of nonlinear oscillation system modeling, key parameter identification, frequency dynamic learning, and online monitoring and early warning to ensure the operability and feasibility of the technical solution.

[0021] Modeling of nonlinear oscillating systems This step is based on the laws of conservation of mass and momentum, and uses the lumped parameter modeling method to complete the dynamic abstraction of the compressor-cavity-throttle valve system. This method simplifies the analysis process of complex flow fields by equating continuously distributed system parameters with lumped parameters, and is suitable for the aerodynamic stability study of gas turbine compressors.

[0022] First, a system of dimensionless governing equations is established. The purpose of dimensionless processing is to eliminate the influence of physical dimensions and unify the analysis standards under different working conditions. The equations are expressed as follows: The physical meaning of each parameter in the formula is clear: Ψ is the dimensionless pressure coefficient at the compressor outlet, and its rate of change is proportional to the dimensionless flow coefficient flowing into the cavity. Dimensionless flow rate through the throttle valve difference; The dimensionless flow coefficient of the compressor, whose rate of change is proportional to the pressure rise provided by the compressor itself. The difference between the back pressure Ψ of the cavity; τ is a dimensionless time, determined by the ratio of actual time to the system characteristic time; B is a stability parameter, a core parameter that determines whether the system experiences surge, and its calculation formula is: Where U is the rotor circumferential velocity. V is the inlet sound velocity, V is the cavity volume, A is the equivalent flow channel area, and L is the equivalent flow channel length. The compressor characteristic curve is a nonlinear function of flow rate and the core source of system nonlinearity. The characteristic curve of the throttle valve is simplified to... The throttling coefficient is determined by the structure and opening degree of the throttling valve.

[0023] Secondly, a small disturbance analysis was performed on the system to select the compressor's stable operating point. This operating point is the flow-pressure matching point of the gas turbine under stable operating conditions. Let... ,in , It is a small disturbance quantity that is much smaller than the steady-state value, which meets the application conditions of small disturbance analysis.

[0024] Compressor characteristics and throttle valve characteristics At the stable operating point, a third-order Taylor expansion is performed, retaining up to the cubic term to capture the nonlinear instability behavior of the system. The expanded expression is then substituted into the dimensionless control equations, and after algebraic transformation and elimination, the quadratic nonlinear term is eliminated to obtain the expression for the disturbance flow rate. The dynamic equations.

[0025] Finally, variable substitution and standardization are performed to make... , and define ,in The coefficients after Taylor expansion are given, resulting in a nonlinear oscillating system model with cubic nonlinear stiffness. This model accurately describes the axial reciprocating oscillation characteristics of the compressor airflow, as shown in the following equations: Key parameter identification To ensure that the nonlinear oscillating system model accurately reflects the dynamic characteristics of the target gas turbine, it is necessary to determine the nonlinear stiffness coefficient η, damping ratio ζ, and natural frequency based on compressor characteristic data and real-time operating data. The online identification process consists of the following five steps: Compressor characteristic data acquisition There are two ways to obtain characteristic data: one is to look up the steady-state characteristic curves of the target compressor across the entire speed range in the gas turbine design manual. Secondly, the results were obtained through bench tests. During the tests, the gas turbine was controlled to run stably at different speeds, the throttle valve opening was adjusted to change the load, and the compressor outlet pressure and flow data were collected simultaneously. The least squares method was used to fit and obtain a high-precision characteristic curve.

[0026] Stable operating point determination The gas turbine speed N and throttle valve opening are acquired in real time using a sampling frequency that matches the system's inherent high frequency. Data. Substitute the collected parameters into the obtained compressor steady-state characteristic curve. The stable operating point under the current operating condition is obtained through interpolation calculation and matching. This ensures the timeliness and accuracy of work at the work sites.

[0027] Compressor characteristic derivative calculation The compressor characteristic curve at the steady operating point is calculated using the central difference method. The first, second, and third derivatives at a given point are calculated using this method, which offers higher accuracy than forward and backward differencing. The specific formulas are as follows: In the formula, Δφ represents a small flow increment, which is taken as the steady-state operating point flow rate. It accounts for 1% to 2%, balancing computational accuracy and efficiency.

[0028] Calculation of linearization coefficient and stability parameters of throttle valve Based on the characteristic curve of the throttle valve Calculate its pressure at the steady operating point. The linearization coefficient β at point is given by the formula: Meanwhile, by combining parameters such as rotor circumferential speed U, inlet sound velocity a, and cavity volume V under the current operating conditions, and substituting them into the calculation formula for stability parameter B, the real-time calculation of the value of B is completed.

[0029] Model key parameter calculation The physical property parameters obtained in the above steps Substituting the parameters into the correlation function, we obtain the key parameters of the nonlinear oscillating system model. The correlation function expression is as follows: The above function is derived from the characteristics of the compressor and the throttle valve, which can achieve accurate mapping between model parameters and physical parameters, ensuring that the model can track the operating conditions of the gas turbine in real time.

[0030] Frequency dynamic learning The core of this step is to construct a frequency-learnable nonlinear oscillation system by dynamically adjusting its natural frequency. This enables the model to adapt to a wide range of operating conditions. The specific implementation process is as follows: Frequency learning mechanism settings The natural frequency in the nonlinear oscillating system model The fixed value is replaced by a dynamic variable, and its update process follows a preset learning rule, as shown in the following expression: In the formula The learning rate should be chosen to balance update speed and model stability, and is generally selected in the range of 0.001 to 0.01. The low-frequency characteristic of the pre-calibrated surge precursor mode is determined by the structure and aerodynamic characteristics of the gas turbine. The inherent high frequency of the system originates from combustion noise inside the gas turbine, the passing frequency of the blades, etc.

[0031] The working principle of learning mechanisms When the operating conditions of the gas turbine change, the excitation signal of the system This will change accordingly; the learning rules will adjust the natural frequency in real time based on changes in the excitation signal. .

[0032] As the rotational speed increases, the frequency of the blades passing through... As the amplitude increases, the amplitude and frequency characteristics of the excitation signal change. This will increase accordingly, ensuring the model always matches the current operating conditions; when the load fluctuates, the system's nonlinear characteristics intensify, and changes in the amplitude of the excitation signal will drive... Make adaptive adjustments to ensure the model's effectiveness across the entire operating range.

[0033] Online monitoring and early warning This step achieves ultra-early warning of gas turbine surge boundary by collecting operating parameters, calculating response amplitude, and establishing a baseline model. It is specifically divided into three stages: Operational parameter acquisition and high-frequency disturbance extraction Using existing pressure, flow, and speed sensors at the compressor outlet, pressure, flow rate, speed, and inlet condition parameters are acquired online at a high sampling rate of no less than 10kHz. The acquired raw data undergoes preprocessing, which includes: bandpass filtering to remove low-frequency drift and high-frequency noise, and wavelet denoising to eliminate environmental interference signals, ensuring data purity.

[0034] Extract high-frequency disturbance information from the preprocessed signal. The extraction method is Fast Fourier Transform (FFT). By performing spectral analysis on the signal, the inherent high-frequency frequencies of the system are located. The corresponding frequency band is used to extract the signal within that frequency band as an auxiliary excitation.

[0035] Calculation of characteristic low-frequency response amplitude Based on a frequency-learnable nonlinear oscillatory system model, the system's response to characteristic low-frequency information is calculated. The response amplitude R is determined by the following steps: Steady-state data extraction: The sliding window method is used to extract the steady-state response time series of the system. The window length is set to 5 to 10 low-frequency characteristic cycles to ensure that the captured data is a stable response and to eliminate interference from the transition process.

[0036] Fourier coefficient calculation: Perform a discrete Fourier transform on the truncated steady-state response time series to calculate the characteristic low-frequency components. The sine Fourier coefficient at the point With cosine Fourier coefficient The formula is as follows: In the formula, N is the number of sampling points in the steady-state response time series. This is the l-th sampling time.

[0037] Response amplitude calculation: and Substituting into the amplitude calculation formula, we obtain the response amplitude R, as shown in the following formula: The response amplitude R is a quantitative indicator that characterizes the stability margin of a system. The larger the R value, the stronger the nonlinearity of the system and the closer it is to the surge boundary.

[0038] Baseline model establishment and graded early warning triggering Baseline Model Establishment: During the healthy and stable operation of the gas turbine, response amplitude R data is collected across the entire operating range, covering all speed ranges from idle to rated speed, and all load ranges from no-load to full-load. A polynomial fitting algorithm is used to establish a correlation model between the R value and operating parameters such as speed and throttle valve opening, i.e., the baseline model; simultaneously, warning thresholds under different operating conditions are determined based on the 3σ principle, where σ is the standard deviation of the R value under healthy conditions. Tiered early warning triggering: The real-time calculated R value is compared with the baseline model, and the early warning level is divided according to the degree of deviation and the rate of increase of the R value. Level 1 warning: The R value exceeds the baseline by 1~2σ, indicating that the system stability margin has slightly decreased, and the operating parameters need to be closely monitored; Level 2 warning: The R value exceeds the baseline by 2~3σ, indicating that the system nonlinearity is enhanced, and the operating parameters need to be actively adjusted (such as reducing the speed and adjusting the throttle valve opening). Level 3 warning: If the R value exceeds the baseline by more than 3σ, it indicates that the system is about to experience surge, and an active stability control strategy should be triggered immediately to prevent surge from occurring.

[0039] Through the complete implementation process described above, high-precision and high-reliability online identification and ultra-early warning of gas turbine compressor surge boundaries can be achieved, providing a guarantee for the safe operation of gas turbines.

[0040] Working principle: Based on nonlinear dynamics theory, the compressor-cavity-throttle valve system is abstracted as a cubic nonlinear stiffness oscillation system with dynamically learnable natural frequencies. The response signal of the low-frequency surge precursor mode is amplified by the system's inherent high-frequency disturbances. By monitoring abnormal changes in the response amplitude, an ultra-early warning of the surge boundary is achieved. The specific logic link is as follows: The system dynamics abstraction is based on the laws of mass conservation and momentum conservation. Through dimensionless transformation and small perturbation analysis, the axial flow perturbation or pressure pulsation of the compressor airflow is transformed into the dynamic variable Θ of the nonlinear oscillating system. Combined with the third-order Taylor expansion of the compressor characteristic curve, a cubic nonlinear stiffness oscillation model that can accurately characterize the nonlinear instability behavior of the system is constructed, thereby achieving mathematical simplification of complex aerodynamic processes.

[0041] The utilization of high-frequency excitation and dynamic frequency adaptation do not require additional external excitation; they directly extract inherent high-frequency disturbances such as combustion noise and blade pass-through frequency from inside the gas turbine as auxiliary excitation. Simultaneously, the model's inherent frequency Designed as a dynamic variable, the model adjusts in real time according to changes in the excitation signal through preset learning rules, enabling it to adapt to a wide range of speeds and load conditions, thus overcoming the application limitations of traditional static models.

[0042] Amplification and Quantitative Monitoring of Low-Frequency Precursors: As the compressor operating point approaches the surge boundary, the system's nonlinear characteristics intensify, and the high-frequency auxiliary excitation interacts with the low-frequency surge precursor modes within the flow channel. This generates a nonlinear coupling effect, causing the response amplitude R of the low-frequency mode to rise rapidly. By calculating this amplitude and establishing a baseline model for all operating conditions, the R value can be used as a quantitative indicator of the system's stability margin.

[0043] The triggering logic of the graded early warning compares the deviation and rate of rise of the R value with the baseline model in real time, and classifies the early warning level according to the degree of deviation: when the deviation is slight, the operating parameter monitoring is initiated; when the deviation is moderate, the operating condition parameters are actively adjusted; and when the deviation is severe, the emergency control strategy is triggered, forming a closed-loop early warning mechanism of "monitoring-assessment-intervention", ultimately achieving ultra-early identification and prevention of compressor surge.

[0044] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for early warning of surge boundary in a gas turbine, characterized in that, Includes the following steps: S1: Based on the laws of conservation of mass and momentum, establish the dimensionless control equations of the compressor-cavity-throttle valve system and construct the nonlinear dynamic model of the compressor system; S2: Fit the compressor characteristic curve, introduce disturbance variables to construct a nonlinear oscillating system model with cubic nonlinear stiffness, and the natural frequency of the nonlinear oscillating system model can be dynamically updated; S3: Based on compressor characteristic data and operating data, identify the nonlinear stiffness coefficient η, damping ratio ζ, and natural frequency of the nonlinear oscillation system model online. ; S4: Collect compressor operating parameters and extract high-frequency disturbance information from the collected parameters. Calculate the nonlinear oscillation system at the characteristic low frequency. The response amplitude R at the location is used to establish a baseline model of how the R value changes with time or operating conditions. When the R value exceeds the baseline threshold and continues to rise, it is determined that the system is approaching the surge boundary and an early warning is issued.

2. The gas turbine surge boundary early warning method according to claim 1, characterized in that, In step S2, the key parameters of the nonlinear oscillating system model with cubic nonlinear stiffness are related to the compressor physical characteristic parameters through the following formula: In the formula, For stability parameters, It is related to rotational speed, cavity volume, and flow channel size; β is the linearization coefficient of the throttle valve; This is the first derivative of the compressor characteristic curve at the steady operating point; This is the second derivative of the compressor characteristic curve at the steady operating point; This is the third derivative of the compressor characteristic curve at the steady operating point; A function characterizing the relationship between natural frequency and compressor physical characteristic parameters; It is a function that characterizes the relationship between the damping ratio and the physical characteristic parameters of the compressor; It is a function that characterizes the relationship between the nonlinear stiffness coefficient and the physical characteristic parameters of the compressor.

3. The gas turbine surge boundary early warning method according to claim 1, characterized in that, In step S2, the specific process of constructing a nonlinear oscillating system model with cubic nonlinear stiffness is as follows: at the stable operating point of the compressor, the compressor characteristic curve is expanded in third order Taylor, and the second-order nonlinear term is eliminated by coordinate transformation to obtain a nonlinear oscillating system model describing the axial reciprocating oscillation characteristics of the compressor airflow.

4. The gas turbine surge boundary early warning method according to claim 1, characterized in that, The parameter identification process in step S3 includes the following stages: First stage: Based on the steady-state characteristic curve data of the target compressor, determine the current stable operating point according to the real-time gas turbine speed and throttle valve opening. ,in To stabilize the dimensionless flow coefficient at the operating point, The dimensionless pressure coefficient for the stable operating point; Second stage: Calculate the compressor characteristic curve at the stable operating point. The first derivative at Second derivative Third derivative The first stage involves calculating the linearization coefficient β of the throttle valve under the current operating conditions and the stability parameter B. The second stage, based on the calculated compressor physical characteristic parameters, involves obtaining the nonlinear stiffness coefficient η, damping ratio ζ, and natural frequency. .

5. The gas turbine surge boundary early warning method according to claim 1, characterized in that, In step S4, the nonlinear oscillation system receives characteristic low-frequency information. The response amplitude R is calculated using the following formula: In the formula, Characteristic low frequency The sine Fourier coefficient at the location; Characteristic low frequency The cosine Fourier coefficient at the point; and By analyzing the steady-state response time series of the system model It can be obtained by integration or discrete summation.

6. The gas turbine surge boundary early warning method according to claim 1, characterized in that, In step S4, the compressor operating parameters include compressor outlet pressure, outlet flow rate, rotor speed, and inlet condition parameters.

7. The gas turbine surge boundary early warning method according to claim 1, characterized in that, In step S4, the high-frequency disturbance information The source is at least one of the following: combustion noise inside the gas turbine system or the passing frequency of the blades.

8. The gas turbine surge boundary early warning method according to claim 5, characterized in that, In step S4, the process of establishing the baseline model is as follows: during the healthy and stable operation phase of the gas turbine, the R value is collected under different speeds and different throttle valve opening conditions, and a correlation model between the R value and the operating parameters is established.

9. A gas turbine surge boundary early warning system for implementing the method of any one of claims 1-8, characterized in that, include: Data acquisition module: Used to collect compressor operating parameters online and extract high-frequency disturbance information from the collected parameters. ; Model identification module: used to derive the nonlinear dynamic model of the compressor system, identify and update the parameters of the nonlinear oscillating system with cubic nonlinear stiffness; Response monitoring module: used to calculate the system response amplitude R online, and perform trend analysis and baseline comparison on the R value; Warning output module: Used to generate graded warning information when the R value exceeds the baseline threshold and continues to rise, and output the warning information to the gas turbine control system.

10. The gas turbine surge boundary early warning system according to claim 9, characterized in that, The system is integrated into the gas turbine controller or a separate health management unit. The system supports linkage with the gas turbine active stability control system. The sampling frequency of the data acquisition module is matched with the inherent high frequency of the compressor.

Images